17 commits

Author	SHA1	Message	Date
ver4a	6a7150ef3e	build: add spirv-headers ... This fixes an issue with new dependency mpv build has: "spirv-headers is required to build with Vulkan support enabled" The commit: 72a0b20e42	2026-04-03 01:19:43 +02:00
ver4a	e57c274511	Revert "build: move configuration after build scripts" ... This change broke scx config generation. This reverts commit 3ec2883d7b.	2026-02-20 10:18:54 +01:00
ver4a	3ec2883d7b	build: move configuration after build scripts ... Removing bash-color-prompt was removing my custom prompt, this restores the previous behavior.	2026-02-19 19:43:49 +01:00
ver4a	dddd72d82a	build: move kde-onc image build logic to bash ... This makes the build significantly faster, as podman/buildah has a large processing overhead for commiting each layer, even though I squash them in the end.	2026-02-13 18:42:45 +01:00
ver4a	ca9a24c308	feat: add scx_bpfland scheduler ... I'm having some issues with scx_lavd not using P-cores so far, so I'm going to experiement with scx_bpfland some more.	2026-01-28 22:05:45 +01:00
ver4a	befe1df9a3	fix: use correct kernel version ... uname reports running/host kernel, we need the kernel contained in the image, that's the one that'll be running on a booted system. This same thing applies to the nvidia layer where I used the same method as I do here now.	2026-01-25 22:40:42 +01:00
ver4a	26be9035bd	build: move kconfig fix-up to correct position in Dockerfile	2026-01-25 21:27:44 +01:00
ver4a	1165bdc3e4	fix: fix-up the kconfig path for csx_lavd ... Previous commit added an environment file with command substitution, which is not possible. This commit fixes that by generating the file during build. This solution should be reasonably robust, because for every image there is only a single kernel version.	2026-01-25 19:08:40 +01:00
ver4a	fbe63af1bf	fix: fix scx_lavd build	2026-01-25 18:07:32 +01:00
ver4a	618b14d444	fix: fix build	2026-01-25 17:41:12 +01:00
ver4a	cadbd05c4a	feat: add scx_lavd scheduler ... It seems to be the most energy efficient on my lunar lake laptop, having the lowest low-utilization power draw (scx_lavd intel_pstate active > intel pstate active > EAS intel pstate passive), but there are some odd performance results, where only E-cores are used for single-threaded workloads like "zstd -b1" and "pv /dev/urandom > /dev/null". From my initial testing it seems to be way more convervative with using P-cores for these specific workloads, and only uses them if most of E-cores get overloaded, unlike EAS that engages P-cores when one E-core is utilized above 80%.	2026-01-25 17:24:14 +01:00
ver4a	cf27ab47ba	feat: load ntsync kernel module on boot ... ntsync was merged into wine and will be included in future proton releases Fedora may enable this by default https://fedoraproject.org/wiki/Changes/NTSYNC-Contained, until that happens, this package is necessary. And may remain necessary because of flatpaks (e.g. steam), that can't depend on rpms and load modules through them.	2026-01-10 20:33:46 +01:00
ver4a	b4ea8af4e0	Rebase to F43 again	2026-01-07 22:53:20 +01:00
ver4a	9a4f51fc5d	Revert "Rebase to F43" ... The nvidia build is failing on F43, because nvidia akmod is not getting installed correctly. I'll need to investigate this before bumping back to F43. This reverts commit 802ccc033a.	2025-12-09 17:48:46 +01:00
ver4a	802ccc033a	Rebase to F43	2025-11-25 16:18:38 +01:00
ver4a	265e4b1e88	(mpv) Add LittleCMS 2	2025-07-30 21:37:06 +02:00
ver4a	c61a0fa2a0	Optimize repository	2025-04-29 15:08:01 +02:00

Renamed from Dockerfile.kde (Browse further)